Websites AI cites most in Energy

Few people ask an AI assistant about a heat pump out of curiosity. Prompts in this industry are about payback, subsidies and risk: “are solar panels worth it in Brittany”, “how much does an air-to-water heat pump cost after grants”, “can I sell my surplus electricity”, “kiln-dried or seasoned logs, which burns better”.

The second large family is about choosing a supplier: “how to find a reliable solar installer”, “which heat pump brand is the quietest”, “firewood delivery near Leeds”. In the answers, the financial context nearly always comes first and the company names come after.

1. Energy agencies and ministries. France Rénov’, the national housing agency ANAH, ADEME and the ecology and economy ministries explain grants, standards and savings. Belgian regions run their own energy portals, cited in the same way.
2. Quality labels. Qualit’EnR and Qualibat appear whenever an answer explains how to check an installer. In England, the Ready to Burn certification run by Woodsure plays a comparable role for firewood.
3. Grid operators and utilities. Enedis, which runs most of the French distribution grid, is cited for grid connection and self-consumption, and large suppliers such as TotalEnergies for contracts and offers.
4. Subsidy and comparison platforms. Services that help households claim energy-saving premiums or compare installers are cited for grant amounts and steps.
5. Consumer associations. UFC-Que Choisir is used for warnings and independent comparisons.
6. Installers and fuel sellers. Regional solar installers, heat pump specialists and log merchants are cited from their own sites, usually when the prompt names a region.

In France, the main public grants for energy renovation are paid only when the work is carried out by a company holding the RGE label (Reconnu Garant de l’Environnement). AI models know this and repeat it, so an answer about a heat pump or a solar installation often tells the user to check the installer in the official directory before signing.

The label then acts as a filter on which companies are worth naming. Qualit’EnR splits its schemes by technology (QualiPV for photovoltaics, QualiPAC for heat pumps, Qualibois for wood heating), and an installer that states the exact scheme it holds gives the model a precise, checkable fact.

The sector has also seen aggressive doorstep selling and fraudulent renovation offers, which public agencies warn against. AI answers relay those warnings, which favours companies whose legal details, certifications and references are public.

In this sector, GEO (Generative Engine Optimization, optimising content so that AI assistants cite it) rests less on the volume of content than on proof: a label number, a measured output, a dated explanation of the grants.

National brands and subsidy platforms answer the general questions, but installers can win the regional ones. A prompt that names a department or a city regularly leads to the websites of local companies, sometimes very small ones, provided they mention that area explicitly.

English questions in this dataset have a different profile: many concern firewood and wood fuel in the United Kingdom, where log suppliers, drying methods and certification are the main topics. The mix of sources in this industry therefore reflects the products sold by the companies analysed as much as the energy market as a whole.